inquisitrlogo

 
Why we need an open access Twitter client


tweetdeck-2

The Twitter application market is booming at the moment off the back of Twitter’s phenomenal growth this year. Seesmic launched Seesmic Desktop yesterday (the new Twhirl), delivering multi-account access to a multi-column client. TweetDeck is out with a new version that promises to have fixed its memory issues, and adds some extra features for good measure. Mac client Nambu is starting to get a lot of attention, and rightly so, it’s quick and offers a non-Ajax alternative. CoTweet doesn’t offer a desktop client yet, but brings CRM to Twitter, marking a significant step forward in the Twitter landscape.

This post was going to start out as a Twitter client review, but Rafe Needleman wrote the post I would have written at CNet yesterday, give or take a few things I would have covered differently (read it here.) The choice we have in Twitter clients has never been better, and yet none are perfect.

The feature problem

The differences between each client come down to features, with no client offering the lot. TweetDeck may have popularized the multi-column layout, but its feature set is also the richest on the market. This may seem like a small thing, and it’s a personal preference, but the feature I love the most on TweetDeck is TweetShrink. This one button allows you to automatically shrink a destroytwitterTweet that has gone over the character limit. And yet TweetDeck doesn’t support multiple Twitter accounts and a range of other services (although it does support 12seconds and Facebook.) Seesmic Desktop rocks with multiple Twitter account support, although doesn’t support other services yet (Twhirl did, and we’re told these are coming.) Seesmic Desktop doesn’t offer the variety of additional Twitter tools TweetDeck does, although does offer image capture by webcam (a nice touch.) Nambu seems to be more a front for URL shortening and picture hosting services owned by the company that wrote it, and you’re locked in. DestroyTwitter, another smart multi-column Twitter client supports skinning, a feature the other clients simply don’t offer.

It’s unreasonable to expect that the creators of each Twitter client could and should offer every single imaginable feature, even beyond the constraint that for most, there’s little to no business model/ revenue from these clients (at least yet.)

But there is an army of people out there who given a platform could provide these features.

Open Twitter client

The solution to the feature problem is an open Twitter client. A Twitter platform might not be the right description, but it goes close.

This Twitter client would offer support for plugins or extensions is a similar way that Firefox does. Users could create plugins that add features to the client. For example, a plugin that added Identi.ca support to TweetDeck, TweetShrink to Seesmic Desktop, or even Flickr image uploading to Nambu. If you get the ground work right, the only limitation is the technical constraints and imagination of your users.

There may be programing issues in offering an open platform in Ajax, for example I don’t recall ever seeing an Ajax program offer plugins, although DestroyTwitter offers skins, suggesting that there is a way to bring in additional content/ services. Nambu is OS X native, so plugin support is possible.

Encouraging competition

As mentioned above, Nambu suffers from lock-in to services owned by the same people who wrote the client. The new version of TweetDeck has actually reduced the options of URL shortening. Most services only offer TwitPic for image uploading. As we increase the diversity of Twitter clients, we’re actually seeing a consolidation in the tie-ins to related services.

An open Twitter client would allow competing services to write their own plugins, giving users a true choice within the client. For example, I’d love to be able to link in to Skitch for images in my Twitter client at the push of a button or two, as opposed to the current need to upload, visit the site, grab the URL, then paste it in. There are a range of 3rd party stats sites now that would make nice value adds; TweetDeck offers Twitscoop, but imagine other services, complete with search (Tweetstats or similar sites.)

This isn’t to say that a service like TwitPic aren’t worthy (it’s a great service) but lock-in reduces competition, and reduced competition usually results in slower innovation. It’s in the communities interest for there to be a vibrant set of related tools accessible via their Twitter client.

Conclusion

The idea of an open Twitter client may to some degree be a pipe dream. We may not see it upfront, but I’d bet that a lot of the services pushed in clients like TweetDeck now aren’t offered from the goodness of the heart, but through business relationships which may or may not include money; if they don’t include money now, they likely soon will, because these 3rd party tools want users, and a button in a Twitter client can possibly deliver users by the million.

On alternative is to go open source, although that may be stretch, but likewise is not impossible.

Delivering an open Twitter client done right could end up being a selling point that defines the client in a sea of competition. Imagine a TweetDeck or Seesmic Desktop with hundreds of plugins and theme options which you can pick and choose from, vs an existing client with a limited range of value adds. That’s got to be an appealing selling point, if someone is willing to take it on.











Comments


3 Archived Responses to “ Why we need an open access Twitter client ”

  1. This is why, somewhat along these lines, I created Spaz two years ago. Of course, it actually predates Tweetdeck, Nambu, and even Twhirl (err, Seesmic Desktop, sorry). But is has from day 1 been fully open-source under a New BSD-style license. Spaz enjoys some popularity, although it lacks the VC funding that helps drive the promotion engine of other clients. There is a small but dedicated team of folks who help out with development and support, and people are *always* welcome to check out the source and submit patches for new features. Most don't, but a few do, and it's how we've gotten support for numerous url shortening and image hosting services.

    More recently I've started the SpazCore project to provide a base set of libraries that work across browser engine platforms like AIR, Titanium, and webOS. We have the same license and policies for submission as the original Spaz codebase. SpazCore will serve as the basis for a full rewrite of Spaz (the current codebase is a bit icky) that will start after version 0.8.0 comes out this week.

    While it may not fit all of your needs — I'd argue that nothing is going to do *everything* you want, unless you write it yourself — a project like you describe exists, and has for a while. You can learn more about it at http://funkatron.com/spaz

  2. Why not take it a step further and work on an open twitterish microblogging protocol. If you're worried about competition, why not bypass the walled garden that is Twitter? While the little bird's API is friendly enough in opening the data, an extensible messaging protocol would allow us to consume both the data streams and functions to perform on the data. Continuing the follower structure keeps the system less susceptible to spam.

  3. I think we must sort out the terminology – Open Access, Open Source and Plugins are often mentioned together but one is not mandatory to the others.

    There are plenty of Closed-Source Projects with support of third party plugins. You can start from Adobe (Photoshop, Illustrator,…) going over Winamp and many more.

    I agree that a Twitter-Client with Plugin Support would be cool, but this doesn't mean it have to be open source.